Phillip D. Shade , Senior Network Engineer · Expert), Cisco CCNA, CWNA (Certified Wireless Network...

Post on 28-Feb-2020

9 views 0 download

Transcript of Phillip D. Shade , Senior Network Engineer · Expert), Cisco CCNA, CWNA (Certified Wireless Network...

Phillip D. Shade , Senior Network Engineer Merlion’s Keep Consulting

1

Phillip D. Shade (Phill) phill.shade@gmail.com

•  Phillip D. Shade is the founder of Merlion’s Keep Consulting, a

professional services company specializing in Network and Forensics Analysis

•  Internationally recognized Network Security and Forensics expert, with over 30 years of experience

•  Member of FBI InfraGard, Computer Security Institute, the IEEE and Volunteer at the Cyber Warfare Forum Initiative

•  Numerous certifications including CNX-Ethernet (Certified Network Expert), Cisco CCNA, CWNA (Certified Wireless Network Administrator), WildPackets PasTech and WNAX (WildPackets Certified Network Forensics Analysis Expert)

•  Certified instructor for a number of advanced Network Training academies including Wireshark University, Global Knowledge, Sniffer University, and Planet-3 Wireless Academy.

Telephony Perceptions Through the Years….

VoIP / Video Protocol Stack

Data Link Layer Protocol

IPv4 / IPv6

TCP UDP

H.245

SDP / SIP

MGCP RTP

RTCP H.225

H.323 Specification Codec -

G.711, G.729 H.261, H.263

Media  Call  Control  &  Signalling  

RAS Q.931 SCCP

Unistem

VoIP Protocols Overview (In-band Signaling)

•  MGCP - Media Gateway Control Protocol –  Defined by the IETF and ITU and used to control signaling and session

management (also known as H.248 or Megaco)

•  SCCP - Skinny Client Control Protocol –  CISCO proprietary protocol used to communicate between a H.323 Proxy (performing H.225

& H.245 signaling) and a Skinny Client (VoIP phone)

•  SIP - Session Initiation Protocol –  Defined by the IETF / RFC 2543 / RFC 3261

•  H.323 – Defines a Suite of ITU designed protocols –  H.225, H.245, Q.931, RAS, etc…

VoIP Protocols Overview (Data)

•  RTP - Real Time Protocol –  Defined by the IETF / RFC 1889 –  Provides end-to-end transport functions for applications transmitting real-time data over

Multicast or Unicast network services (Audio, video or simulation data)

•  RTCP - Real Time Control Protocol –  Defined by the IETF –  Supplements RTP’s data transport to allow monitoring of the data delivery in a manner

scalable to large Multicast networks to provides minimal control and identification functionality

•  RTSP - Real Time Streaming Protocol –  Defined by the IETF / RFC 2326 –  Enables the controlled delivery of real-time data, such as audio and video; designed to

work with established protocols, such as RTP and HTTP

Codecs (Audio / Video Conversion)

•  CODEC = Compressor / Decompressor or Coder / Decoder or Reader - Provides conversion between Audio/Video signals and data streams at various rates and delays

•  Designations conform to the relevant ITU standard –  Audio Codecs (G.7xx series)

•  G.711a / u - PCM Audio 56 and 64 Kbps (Most common business use) •  G.722 - 7 Khz Audio at 48, 56 and 64 Kbps •  G.723.1 / 2- ACELP Speech at 5.3 Kbps / MPMLQ at 6.3 Kbps •  G.726 - ADPCM Speech at 16, 24, 32 and 40 Kbps •  G.727 - E-ADPCM Speech at 16, 24, 32 and 40 Kbps •  G.728 - LD-CELP Speech at 16 Kbps •  G.729 - CS-ACELP Speech at 8 and 13 Kbps (Very common for home use)

–  Video Codecs (H.2xx series) •  H.261 - Video >= 64 Kbps •  H.263 / H.264 - Video <= 64 Kbps

Analog in Digital conversion via Codec Analog out

•   MOS  and  R  value  include  Packe8az8on  delay  +  Ji=er  buffer  delay  •   Common  bandwidth  –  real  bandwidth  consump8on:      #  Payload  =  20  bytes/p  (40  bytes/s)      #  Overhead  includes  40  bytes  of  RTP  header  (20  IP  +  8  UDP  +  12  RTP)  

Sample VoIP Codec Comparison

Competing In-Band Signaling Standards

•  Several different standards are currently competing for dominance in the VoIP field: –  H.323 - Developed by the International Telecommunications Union (ITU) and the

Internet Engineering Task Force (IETF)

–  MGCP / Megaco/ H.248 - Developed by CISCO as an alternative to H.323

–  SIP - Developed by 3Com as an alternative to H.323

–  SCCP – Cisco Skinny Client Control Protocol – used to communicate between a H.323 Proxy (performing H.225 & H.245 signaling) and a Skinny Client (VoIP phone)

–  UNISTEM – Proprietary Nortel protocol, developed by as an alternative to H.323

H.323 - Packet-based Multimedia Communications Systems

•  An umbrella standard defined by the International Telecommunications Union (ITU) and the Internet Engineering Task Force (IETF)

•  Defines a set of call controls, channel set up and Codec’s for multimedia, packet-based communications systems using IP-based networks

H.450.1 Supplemental, generic protocol for use under H.323

H.225 Call Signaling / RAS

H.245 Control messages for the H.323 Terminal (RTP / RTCP)

H.235 Security Enhancements Q.931 Call setup and termination

G.711, G.723.1 G.728 Audio Codec's

H.261, H.263, H.264 Video Codec’s

VoIP Standard (SIP)

•  Defined in RFC 2543 and RFC 3261 and by the ITU –  Pioneered by 3Com to address weaknesses in H.323

•  Application layer signaling protocol supporting real time calls and conferences (often involving multiple users) over IP networks –  Run over UDP / TCP Port 5060 (default) –  Can replace or complement MGCP

•  SIP provides Session Control and the ability to discover remote users •  SDP provides information about the call •  MGCP/SGCP Provides Device Control •  ASCII text based •  Provides a simplified set of response codes

•  Integrated into many Internet-based technologies such as web, email, and directory services such as LDAP and DNS –  Extensively used across WANs

MGCP / Megaco VoIP Standards

•  Defined by RFC 2705 / 3015 and the ITU in conjunction with the H.248 standard –  Pioneered by CISCO to address weaknesses in H.323

•  Used between elements of distributed Gateways (defined later) as opposed to the older, single all-inclusive Gateway device –  Extensively used in the LAN environment

•  Utilizes Media Gateway Control Protocol (MGCP) to control these distributed elements –  Often considered a “Master/Slave” protocol

Quality Of Service (QoS) - Overview

•  Provides a guarantee of bandwidth and availability for requesting applications –  Used to overcome the hostile IP network environment and provide an

acceptable Quality of Service •  Delay, Jitter, Echo, Congestion, Packet loss and Out of Sequence

packets

–  Mean Opinion Score (MoS) / R-Factor is sometimes used to determine the requirements for QoS.

–  Utilized in the VoIP environment in one of several methods: •  Resource Reservation Protocol (RSVP) defined by IETF •  IP Differentiated Services •  IEEE 802.1p and IEEE 802.1q

Assessing Voice Quality

•  Voice Quality can be measured using several criteria 1. Delay: As delay increases, callers begin talking over each other, eventually the call will sound like talking on a “walkie-talkie”. (Over…) 2. Jitter: As jitter increases, the gateway becomes unable to correctly order the packets and the conversation will begin to sound choppy (Some devices utilize jitter buffer technology to compensate) 3. Packet Loss: If packet loss is greater than the jitter buffer, the caller will hear dead air space and the call will sound choppy (Gateways are designed to conceal minor packet loss )

High quality voice connections require all three to be minimized

•  MoS – Mean Opinion Score - Numerical measure of the quality of human speech at the destination end of the circuit

•  PSQM (ITU P.861)/PSQM+ - Perceptual Speech Quality Measure

•  PESQ (ITU P.862) – Perceptual Evaluation of Speech Quality

•  PAMS (British Telecom) Perceptual Analysis Measurement System

•  The E-Model (ITU G.107) – (R-Factor) - Send a signal through the network, and measure the other end!

Different VoIP Quality Measurement Terms

Measures of Voice Quality

•  MOS can only be measured by humans •  R-value can be calculated in software •  PMOS values can be determined from R-value

E-­‐Model  “R”  Factor  scores  comparison  to  MOS  score  

MOS (Mean Opinion Score)

MOS Quality Rating

5 Excellent

4 Good

3 Fair

2 Poor

1 Bad

MOS  -­‐  Mean  Opinion  Score    -­‐  Numerical  measure  of  the  quality  of  human  speech  at  the  des8na8on  end  of  the  circuit  (affected  extensively  by  Ji=er)  -­‐   Uses  subjec8ve  tests  (opinionated  scores)  that  are  mathema8cally  averaged  to  obtain  a  quan8ta8ve  indicator  of  the  system  performance  -­‐  Ra8ng  of  5.0  is  considered  perfect  

E-Model (R-Factor)

•  The E-Model - Recommendation ITU G.107 –  The "E-Model" is a parameter based algorithm based on subjective test results

of auditory tests done in the past compared with current “system parameters”

–  Provides a prediction of the expected quality, as perceived by the user

–  The result of the E-Model calculation is “E-Model Rating R” (0 - 100) which can be transformed to “Predicted MOS (PMOS)” (1 – 5; 5 is non-extended, non-compressed)

•  Typical range for R factors is 50-94 for narrowband telephony and 50-100 for wideband telephony

Cascade  Pilot  Computes  the  R-­‐Factor  and  MOS  scores  

Cascade Pilot – Quality Metrics

Average / Maximum Jitter / Delta and Average / Maximum R-Factor / MOS

Making the Call - Basic VoIP Signal Flow

VoIP Protocol

Signaling

Media

Teardown

Endpoint  #1   Endpoint  #2  

GateKeeper    /  Call  Client  Manager  

Expected SIP Operation

•  To initiate a session –  Caller sends a request to a callee's address in the form of a ASCII text command

•  “Invite” –  Gatekeeper/Gateway attempts phnoe number -> IP mapping/resolution

•  Trying / Response code = 100 •  Ringing / response code = 180

–  Callee responds with an acceptance or rejection of the invitation •  “Accept” / response code=200 “OK”

–  Call process is often mediated by a proxy server or a redirect server for routing purposes

•  To terminate a session –  Either side issues a quit command in ASCII text form

•  “Bye”

Session Initiation Protocol (SIP - Invite)

SIP  is  data  is  carried  in  text  format  

SIP  “Invite”  

Session Initiation Protocol (SIP - Bye)

SIP  -­‐  “Bye”  

Challenges of VoIP

•  Minimize Delay, Jitter and data loss –  Excessive Delay variations can lead to unacceptable data lost or distortion

•  Implementing QoS –  RSVP designed to reserve required resources for VoIP traffic

•  Interoperability of equipment beyond the Intranet –  Different vendors Gateways utilize different Codec’s

•  Compatibility with the PSTN –  Seamless integration required to support services such as smart card and 800 service

Factors Affecting Delay and VoIP Quality - 1

•  Latency –  Round trip latency is the key factor in a call having an “interactive feel” –  <100 msec is considered idle

•  Jitter –  Occurs when packets do not arrive at a constant rate that exceeds the buffering ability

of the receiving device to compensate for –  If excessive Jitter occurs, larger Jitter buffers will be required which cause longer

latency

•  Packet Loss –  Loss of > 10% (non-consecutive packets) will be perceived as a bad connection

Latency Jitter Buffer Latency

Factors Affecting Delay and VoIP Quality - 2

•  Codec Choice - Higher quality = added delay –  Greater the compression factors result in lowered quality - Processing / Encoding /

Decoding

•  Bandwidth Utilization - Less utilization = lower latency, jitter and loss due to collisions

•  Priority - Voice is extremely sensitive to delay –  QoS is used to allow network devices to handle VoIP ahead of other traffic

Voice Quality & Delay

Delay Target

Delay (msec)

800

700

600

500

400

300

200

100 0

High Quality

Satellite Quality

Fax Relay Broadcast Quality

Many factors that contribute to the overall delay are fixed: -Codec delay - Hardware delay - Processing delay - Network physical delay

However, several delay factors are variable: - Queuing delay - Network propagation delay

It is the sum of all of these factors that determines overall delay as shown in the chart to the left

VoIP Delay Calculation Example

IP or WAN Network

End-to-End Delay Not to Exceed 250ms

Compression 20ms

Inter-process 10ms

Decompression 10ms Fixed

Delays

Variable Delays

Queuing 10-20ms

Variable Network Delay: Private IP: determinable

Internet 50-400+ms

Total Fixed Delays (w/o buffer) 71-129ms

Inter-process 10ms

Transmission .25 @ T1

7ms @ 56k

Transmission .25 @ T1

7ms @ 56k

Network (FR) 20-40ms

Buffer Configurable

The #1 Result of Excessive Delay - Jitter

•  Occurs when packets do not arrive at a constant rate that exceeds the buffering ability of the receiving device to compensate for –  Symptoms

•  Often noticed as garbles or a annoying screech during a conversation –  Typical Causes

•  Insufficient bandwidth for the conversation •  Excessive number of Hops in the signal path •  QoS disabled or not supported by one or more devices

VoIP  Packets  leave  at  constant  intervals   VoIP  Packets  arrive  at  variable  intervals  

*

Gateway Gateway

User Symptoms

•  Customer Reported Symptoms –  Cannot place or receive calls –  Hear foreign voices not supposed to be on call (Cross-Talk) –  Volume noticeably low or high –  Choppy Audio –  Features do not work properly

•  Equipment Alarm Indications –  Ring Pre-trip Test Fails –  Internal indications (card, power, etc) –  Loss of Signal / High Error Rate –  Connectivity failures

Analysis of Telephony Protocols - Wireshark

Wireshark has the ability to reconstruct not only VoIP conversations, but also other media streams for later analysis.

Packet Capture File

This example contains four (4) calls and is from a VoIP network using Cisco phones and SIP signaling with G.711 audio codec

VoIP Call Detection, Analysis and Playback